The code for the paper : Distilling Reasoning Capabilities into Smaller Language Models (Coming soon)